12V 20A PSU Crowbar Cct

[Sparky67]

Hi all,

I have been looking through various websites for a cct with component values for an overvolts SCR crowbar (operating at say 15V) for a typical 12V 20A analogue ham radio PSU. There seems to be lots of discussion about them but so far I haven't found anything I can build and incorporate into my PSUs. Does anyone have a suitable design please?

Many thanks.

Martin
G4NCE

[Bazz4CQJ]

I built a crowbar external attachment to use with a commercial linear PSU which I use for my ICOM IC718, circuit attached. Sorry it's tatty, not intended for publication.

There's a Transil diode for fast protection. The two zener diodes (ZD) in series were just hand-picked to add up to the voltage that I wanted to trip at. Seems to do the job.

B

[Sparky67]

Thanks Barrie, I will look at building that design. From a quick bit of research I did it looks like a blade fuse may the way to go to carry the required current OK and to open quickly to protect the external device when the SCR triggers.
Cheers.
Martin

[Radio Wrangler]

That should do the job, but you probably should try it out somehow, just to feel confident it will save your bacon if it ever has to.

Motorola did a crowbar trigger IC that you programmed with the ratio of two resistor values, but I can't remember the type number. It's academic if they aren't available anyway.

If you're designing a complete power supply, you can integrate the crowbar into it in ways that make it a lot less dramatic. In the one I did, I have the crowbar do two things. It opens the relay used for mains on/off, and it uses its thyristor to dump the charge in the reservoir capacitors directly. If I trigger it, there is a little click and I hear the mains switch relay release. It does mean that this supply needs a push on a start button to turn it on, but the whole lot integrates with a soft-start arrangement. When I next come across the circuit diagrams, I'll post them in this group.

David

[turretslug]

The Motorola crowbar IC was probably the MC3423- I acquired a few from a big work-place clear-out long ago (wasn't the inefficiency and management change in British industry occasionally useful, ill wind and all that....) As David says, very simple to make a safety crowbar and if you want further sophistication, the application note is your oyster. It can be made into a postage-stamp matrix-board module light enough to be mounted by its thyristor tab- small enough for a 16V one to go inside my FRG-100 (very low down on the true enthusiast's pecking-order of radios, but expensive and useful enough to me!) and a 19V one inside the TF2002B (worryingly generous regulator headroom, besides I'd sorted out about a dozen different faults on acquisition and didn't want to spend any more time than I could avoid in its innards!)

Colin

[KeithsTV]

That would be the MC3423 crowbar IC. It senses the output voltage and if over the limit triggers a thyristor to short the output. However you need a thyristor with a suitable current rating to discharge the reservoir capacitors.

However it's all academic as David says as the IC is now obsolete.

Keith

[Sparky67]

Plan A is to build and test Barrie's suggested cct on my dual-output bench PSU and then incorporate it into the first (cheapest!) 20A PSU and soak-test it there for a while with the normal levels of on/off and Tx/Rx switching and RF etc in the shack.

I am also considering incorporating a relay soft-start cct into both PSUs, or maybe a 100A pre-built current limiter available from China, or whether a simple NTC thermistor would do the job...?

Martin
G4NCE

[Steve G4WCS]

the PW marchwood was regarded as a decent design and the article describes the various protection circuits used. Practical wireless June 1983 ( scanned in americanradiohistory)

I have PDF copies but they are too large to upload, sorry

[ex 2 Base]

I built the Marchwood PSU and it worked well, with soft start and crowbar, can't believe it was 37 years ago. Ted

[Steve G4WCS]

another useful article here .

http://warc.org.uk/?page_id=404

[duncanlowe]

I thought I recognised this. Many (many) years ago when I was in my first job, one of the other guys in the office was into amateur radio, but wasn't really into electronics construction. He had tried to build this design, but it didn't work. He asked me to look at it and after a bit of work (poorly soldered connections mostly) I got it working for him.

Some years later I inherited a rather dead heavy duty BNOS power supply with a pair of nice torroidals, a big case and space to mount a decent number of 2N3055s. I dug out the photocopy of the article from AR magazine he had given me to help fix his, and built it up. It can supply more than the 20A of the original design, and has adjustable voltage and current limit. And it's still working! Have to admit I've been thinking of getting rid of it, as it's so heavy and I have a nice Farnell AP 60-50 that's smaller, lighter and more powerful (as it's a switcher).

EDIT: oh and I've also found that original photocopy from more than 30 years ago!

[Sparky67]

I finally settled on this cct for the crowbar on the Icom PSU. It has been tested OK in situ with the new fuse removed and 15V applied to the fuse output feeding the circuit.

The 15A blade fuse has been run successfully on my 100W HF SSB transmitter. These fuses carry in excess of the rated current for short periods and I thought it safer to try a 15A fuse here with the 20A fuse feeding the main pass transistors. If it does fail it is easy to replace.

The unit was built on a small piece of stripboard and mounted under an existing screw holding the reg control board in place. The SCR is bolted to the main chassis.

While the PSU was open I took the opportunity to add a mains switch (it was originally controlled by the on/off switch on the radio it was connected to...) and a green '12V OK' LED. The red LED is original and *may* be wired with a zener in series to indicate over-volts at the reg output at some stage.

[Bazz4CQJ]

Hi Martin,
I'm just trying identify what are the key differences (improvements) between that circuit (which I have stored to file) and the slightly simpler one (very conventional) which I had used?

I'm guessing that the TL431 does better job than a zener diode, particularly wrt temperature. Anything else that I'm missing?

I haven't used my Icom for a couple of years (I'm playing with my Trio hybrid), so the crowbar here is not seeing much action.

B

[Terry_VK5TM]

The circuit using the TL431 is more voltage stable with respect to temperature, however, some of this advantage is lost with the inclusion the extra transistor in Sparky67's circuit.

The 100n capacitor on the base of the BC557 also slows down the response time of the circuit considerably (to the human eye, you won't see any difference, but it makes a big difference to solid state devices).

I do make note that the BC557 may be required, depending on the trigger current of the SCR being used, so it is all a trade off against conflicting requirements.

Bottom line is, if it works, that's all that matters.

The Texas instruments datasheet for the TL431 has a crowbar circuit in the applications section.

[jimmc101]

It would have been better if the 47k resistor was returned to the +ve rail rather than the anode of the TL431, this would have removed the Vbe of the BC557 from the equation.
Resistor value change from 47k to 51k should keep the same trip point.

Figure 21 of the TI data sheet for the TL431 shows a simple crowbar circuit using a triac in place of the SCR to remove the need for the PNP transistor.

With all these circuits the SCR / Triac is 'soft' fired ie the gate drive is slowly increased until the device triggered, this can cause hot-spotting leading to failure.
Still safe as the failure is to short circuit, but if repeated tripping is anticipated some form of regenerative drive is advisable.

Jim

[Sparky67]

All noted and thanks for the comments and suggestions. I am going to build a second board for a second PSU so will look at changing the 47k and taking it up to the +ve rail. I actually bought some zeners to series-up for detecting the trigger voltage, but decided to opt for the TL431 later. Just a personal choice. The ‘delay’ cap was the suggested one to stop false triggering so it can always be reduced with experience over time. Better than starting low and increasing it each time the SCR conducts I guess.

As an aside I once had the job of destroying access to some encryption circuitry in a large number of radios so they could be used elsewhere. The agreed quick method after trying it successfully on a few trial sets was to disable the encryption in software and then apply a high voltage to the programming pins to damage the chip so it could no longer be accessed. I was surprised how long it took 30V to do the job. I’m sure some components in a nominal 12V radio will be seen off by a very short application of 20V though...

Thanks again.

Cheers,

Martin
G4NCE

[G6Tanuki]

Take care to RF-proof your crowbar: I've seen a few such circuits which happily handle DC loads but which go into an unwarranted panic-lockdown when there's a hint of RF flying around.

Sprinkle ferrite-beads over your design like confetti; 100-Ohm resistors and 0.01uF capacitors by the handful too.

A 'sensitive' crowbar can also highlight poor slew-rate/loop-gain time-constants in the regulator when dealing with the typical widely-varying load presented by a SSB transceiver: it's all too easy to get 'overswing' transient excessive-voltage at the output of a regulator, which can trip a sensitive crowbar. Designing/implementing a robust regulator to supply a powerful SSB transceiver soon leads you into a realm more-familiar to audio amplifier designers !

[bigfathairyvika]

Google 'PW marchwood'

The first link is a pdf of the Practical wireless 30 amp psu.
Lots of protection explained and provided therein.
Uses the MC3423 crowbar ic.

Mark

[Radio Wrangler]

A friend bought a full HF seup in the late eighties. Ten Tec corsairII, BNOS 30 amp supply, and a big Capco ATU.

Running SSB, the light in the S meter brightened with the speech. Tuning up on CW full power, it went VERY bright.

The Capco wanted a low current 12v supply for relays. Kenneth took this from 12v out phono (Phono? oh boy) on the Ten Tec.

The capco was leaking immense RF voltages back out the 12v input socket... no real filtering. The ten tec had no decoupling on its 12v input and there was large amounts of RF on the transceiver's internal 12v bus. This was probably enough for some brightening, but the RF was reaching the PSU and sending it bananas.

Oh, what a mess!

Good hygiene hardly needs mentioning nowadays, but it is needed with RF too.

David

[Bazz4CQJ]

I must admit that Tanuki's warning has caught me out and that before I use my Crowbar again, I am probably going to make a few changes.

However, my old Kenwood-Trio hybrid is centre-stage at the moment, so the Crowbar is not in use.

B